Testing the most stringent criteria for selection of candidates for active surveillance in patients with low-risk prostate cancer

Authors


Nazareno Suardi, Department of Urology, University Vita-Salute San Raffaele, Via Olgettina 60, 20132 Milano, Italy.
e-mail: suardi.nazareno@hsr.it

Abstract

Study Type – Prognosis (case series)
Level of Evidence 4

OBJECTIVE

To test the ability of two of the most stringent criteria used to identify patients with low-risk prostate cancer suitable for active surveillance (AS) to correctly exclude patients with unfavourable prostate cancer characteristics.

PATIENTS AND METHODS

The study included 874 consecutive patients treated with radical prostatectomy (RP). We selected patients who could have been selected for AS according to the van den Bergh et al. and the Carter et al. criteria. We analysed the rates of advanced disease in these patients, defined as presence of either extracapsular extension (ECE), seminal vesicle invasion (SVI), lymph node invasion (LNI) and Gleason sum of 8–10 or 7–10.

RESULTS

Of 874 patients, 85 (9.7%) and 61 (6.9%) patients, respectively, qualified for AS according to the tested criteria. Within the van den Bergh et al. candidates, 5.9, 1.2, 1.2 and 1.2% of patients, respectively, showed ECE, SVI, LNI and high-grade Gleason sum 8–10 at pathology. Within the Carter et al. candidates, 3.3, 0, 3.3 and 0% of patients, respectively, showed ECE, SVI, LNI and high-grade Gleason sum 8–10. The cumulative rate of unfavourable characteristics was 7.1 and 3.3%. The rate increased to 28.2 and 27.9%, respectively, when Gleason sum 7 was considered as an unfavourable prostate cancer.

CONCLUSIONS

The use of the strictest criteria for AS inclusion identified 7–10% of the men in our cohort of men undergoing RP, as men that would have been eligible for AS. Among this small proportion, between 3.3 and 7.1% of patients harboured unfavourable prostate cancer characteristics. The clinical implications of these misclassification rates remain to be determined.

Abbreviations
AS

active surveillance RP, radical prostatectomy

ECE

extracapsular extension

SVI

seminal vesicle invasion

LNI

lymph node invasion

CaPSURE

the Cancer of the Prostate Strategic Urological Research Endeavor.

INTRODUCTION

Prostate cancer represents the most frequent neoplasm diagnosed in males in the USA and the incidence of prostate cancer diagnosis has been steadily increasing in the last decade [1]. Based on prostate cancer cases diagnosed between 1996 and 2002, an estimated 91% of these new cases will be diagnosed at local or regional stages. In these patients, the 5-year survival is virtually 100%[1]. Therefore, the issue of over treatment of patients with low-risk prostate cancer has been raised. Active surveillance (AS) was proposed as a reasonable option for patients diagnosed with low-risk disease. The aim of AS is to decrease the rate of definitive therapy and its side-effects in patients whose biological tumour characteristics pose a minimal threat to their life expectancy. Several series described the natural history of patients subjected to AS [2–6]. To date, the results associated with this option are controversial: a large proportion of patients submitted to AS show very good oncological outcome [3,4,7,8]. However in these series a consistent proportion of patients had disease progression and needed to be treated after an initial period of observation [3,5,9–11]. Moreover, when patients fail the AS programme and are subjected to radical prostatectomy (RP), high rates of unfavourable prostate cancer characteristics, such as extraprostatic extension (ECE) and high Gleason score, are found at pathological evaluation, given the low volume and low-grade disease that these patients show at the diagnosis [4,12]. These elevated rates of advanced prostate disease may in part be explained by misclassification of patients with unfavourable prostate cancer characteristics among AS candidates. Van den Bergh et al.[5] and Carter et al.[6,7] proposed the most stringent criteria for the identification of patients with low-risk prostate cancer to be included in the AS protocol. In the present study, we tested the ability of these two AS inclusion schemes to correctly exclude patients with unfavourable prostate cancer characteristics.

PATIENTS AND METHODS

We reviewed the contemporary criteria used for identification of AS candidates. Among these, we selected the two most stringent definitions and tested their ability to correctly exclude patients with unfavourable pathological characteristics who would not have been not suitable candidates for AS in our population of patients treated with RP.

For the purpose of the study, we evaluated 874 patients treated with RP at one institution between January 2002 and September 2008 who had their biopsy and RP specimens evaluated at our institution. All patients had complete and detailed clinical and pathological information. Among these patients, we identified those who fulfilled the inclusion criteria for the AS according to the criteria defined by Carter et al.[6,7] and van den Bergh et al.[5]. The criteria used in the Carter et al.[6,7] series were originally proposed by Epstein et al.[13] to define indolent prostate cancer. Carter et al.[6,7] firstly used these criteria for the selection of patients to be included in the AS programme [6]. Table 1[5–7] shows the inclusion parameters for AS according to these criteria.

Table 1.  Selection criteria for AS evaluated in the study and the number of patients that could have been selected for AS according to each variable
VariableCarter et al.[6,7]van den Bergh et al.[5]
Threshold for inclusion in ASN (%)Threshold for inclusion in ASN (%)
  • *

    and <50% of cancer in any core.

Clinical stageT1c524 (59.9)T1c/T2804 (91.9)
PSA level, ng/mL874 (100.0)≤10.0657 (75.2)
Biopsy Gleason score≤3 + 3 = 6596 (68.2)≤3 + 3 = 6596 (68.2)
PSA density, ng/mL per mL≤0.15501 (57.3)<0.2610 (69.8)
No. of positive cores2*131 (14.9)2145 (16.6)
All variables  61 (6.9)  85 (9.7)

Clinical stage was assigned before RP by the attending urologist according to the 2002 TNM staging system. The PSA level was always measured before DRE with the Elecsys Roche Assay. Prostate volume was measured by planimetric calculation during TRUS. Prostatic biopsies were taken with ≥12 cores, with the exception of patients with clinical stage T3 (n = 70; 8.0%) and of patients with PSA value of >20 ng/mL (n = 54; 6.2%), in whom six-core biopsies were taken. All biopsy and RP specimens were examined by two dedicated genitourinary pathologists. All RP specimens were processed according to the Stanford protocol [14].

Statistical analyses consisted of descriptive and categorical analyses where the rates of pathological unfavourable prostate cancer characteristics, defined as the presence of ECE, seminal vesicle invasion (SVI), lymph node invasion (LNI) or Gleason score 8–10 in patients who initially fulfilled the AS criteria. A second set of analyses focused on the presence of ECE, SVI, LNI or Gleason score 7–10.

RESULTS

Table 2[5–7] shows the clinical characteristics of the 874 patients included in the present study. Most patients had cT1c disease at diagnosis (60.0%). The median PSA level was 6.7 ng/mL and the median PSA density was 0.13 ng/mL per mL. The median (range) number of cores taken at prostatic biopsy was 14 (6–29). Only 77 patients (8.8%) had <12 cores taken at biopsy. Most patients had a biopsy Gleason score ≤6 (68.2%). Of 874 patients included in the present study, 61 (6.9%) and 85 (9.7%) patients fulfilled the AS inclusion criteria, respectively, defined by Carter et al.[6,7] and van den Bergh et al.[5] (Table 1) [5–7].

Table 2.  Preoperative characteristics of all patients treated with RP and of patients selected for AS according to each criterion
VariableAll patients treated with RPPatients who would have fulfilled the AS criteria according to:
Carter et al.[6,7]van den Bergh et al.[5]
Overall, n (%)874 (100.0)61 (6.9)85 (9.7)
Mean (median; range):
 Age, years 65.8 (66.3; 44–82)64.7 (66.1; 48–77)64.8 (65.3; 48–81)
 PSA level, ng/mL  9.59 (6.69; 0.2–678.0) 6.58 (6.07; 1.1–17.9) 5.88 (6.0; 1.18–10.0)
 PSA density, ng/mL per mL  0.21 (0.13; 0.02–11.3) 0.09 (0.09; 0.03–0.15) 0.1 (0.09; 0.03–0.19)
 No. of biopsy cores 15.7 (14; 6–29)12.7 (12; 6–20)12.2 (12; 6–20)
 No. of positive cores  6.5 (5; 1–26) 1.8 (2; 1–2) 1.8 (2; 1–2)
N (%):
 Clinical stage
  T1c524 (60)61 (100)63 (74.1)
  T2280 (32) 022 (25.9)
  T3 70 (8.0) 0 0
 Biopsy Gleason score
  2–6596 (68.2)61 (100)85 (100)
  3 + 4185 (21.2) 0 0
  4 + 3 49 (5.6) 0 0
  8–10 44 (5.0) 0 0

Table 1 also shows the number and the proportions of patients that could have qualified for inclusion in AS according to clinical stage, PSA level, biopsy Gleason score, PSA density and number of positive cores [5–7]. For example, one third of contemporary patients cannot be considered for AS due to the presence of a biopsy Gleason score >6. Number of positive cores ≤2 represents the most rate limiting variable for AS. In the present study, the consideration of the number of positive cores allowed the inclusion of only 16.6% of patients according to van den Bergh et al. and of only 14.9% according to Carter et al., which also considers the percentage of cancer in the involved cores. It is noteworthy that each variable (clinical stage, PSA level, PSA density) allows the inclusion of consistent proportions of patients, but when all the variables are considered, very few patients show all the characteristics required by the criteria (6.9% and 9.7% according to Carter et al. and van den Bergh et al., respectively).

Table 3 shows the rate of unfavourable pathological characteristics for patients selected according to the two sets of AS inclusion criteria [5–7]. According to the Carter et al. criteria, the rates of ECE, SVI, LNI and high-grade Gleason sum (8–10) were recorded in, respectively, 3.3, 0, 3.3 and 0% of patients that fulfilled these criteria for inclusion in AS. According to the van den Bergh et al. criteria, ECE, SVI, LNI and high-grade Gleason sum (8–10) were recorded in 5.9, 1.2, 1.2 and 1.2% of patients, respectively. The cumulative rate of unfavourable disease characteristics in patients who fulfilled the inclusion criteria for AS was 3.3 and 7.1%, respectively, when the Carter et al. and the van den Bergh et al. definition was applied. When the presence of Gleason sum 7 was also considered among unfavourable disease characteristics, the rates of pathologically adverse findings increased from 3.3 to 27.9% for the Carter et al. definition and from 7.1 to 28.2% for the van den Bergh et al. definition.

Table 3.  Rates of unfavourable prostate cancer characteristics in patients that could have been selected for AS according to the evaluated criteria
VariableCarter et al.[6,7]van den Bergh et al.[5]
N (%):
ECE 2 (3.3) 5 (5.9)
SVI 0 (0.0) 1 (1.2)
LNI 2 (3.3) 1 (1.2)
Gleason score ≥716 (26.2)24 (28.2)
Gleason score ≥8 0 (0.0) 1 (1.2)
Overall unfavourable (Gleason 7–10)17 (27.9)24 (28.2)
Overall unfavourable (Gleason 8–10) 2 (3.3) 6 (7.1)

DISCUSSION

AS represents a management option for patients with low-risk prostate cancer [4]. The implementation of AS was introduced by Choo et al. in 2002 [15]. Initially, the AS criteria allowed the inclusion of up to 80% of patients with localized prostate cancer, since with the inclusion of biopsy Gleason score up to 7 and of clinical stage up to T2a were permitted [2,9,15]. Subsequently, more strict criteria were proposed. Despite more restrictive criteria, elevated rates of advanced prostate cancer were reported in men followed with AS [16]. More importantly, high rates of patients had disease progression which led to salvage therapies [3,4,9]. For example, 63% of patients who underwent RP in one of the AS protocols showed either extraprostatic disease or LNI [4]. Based on these worrisome data, in a previous analysis on patients who underwent RP, we tested the most popular criteria for AS in 4885 patients from two European institutions [16]. We found that those criteria allowed the inclusion of a significant number of patients into the AS programme, but we showed that up to 26.7% of patients included according to those criteria showed advanced disease (defined as presence of ECE, SVI, LNI or pathological Gleason sum 8/10) at RP [16]. These patients are at high risk of progression, and therefore should not be included in AS protocols.

Based on the limitations of the AS inclusion criteria, more selective definitions of AS criteria were developed [17–19]. In the present study, we decided to test two of the most stringent criteria. One of these definitions originated from the USA in 1994 (Epstein et al.[13]) and was used for the first time to select patients for the AS programme by Carter et al.[6], while the other was proposed by European investigators (van den Bergh et al.[5]). Both sets of criteria have similarities. For example, both rely on the maximum of two positive biopsy cores and on the maximum Gleason score of 3 + 3 at biopsy. Unlike some more historic criteria, the contemporary criteria also rely on PSA density to restrict the inclusion of patients with more aggressive disease characteristics.

The application of the two sets of criteria to the present patient cohort treated with RP showed three important findings. First, the proportion of patients that were eligible for AS was very low. The application of the Carter et al. criteria would only have allowed the inclusion of 6.9% of patients, while application of the van den Bergh et al. criteria would have allowed the inclusion of 9.7% of patients who ultimately underwent RP at our institution. Second, when unfavourable prostate cancer is defined by the presence of ECE, SVI, LNI and Gleason score 8–10 the misclassification rates are low when applying the criteria proposed by Carter et al. and van den Bergh et al. (3.3 and 7.1%, respectively). When comparing the present new results to those previously obtained with less stringent criteria [16], the inclusion of both PSA density and the number of positive cores seems to help in the identification of low-risk prostate cancer, significantly reducing the number of patients with both high-grade (Gleason 8–10) and locally advanced (pT3a/b) disease at pathological examination. Third, still a considerable proportion of patients had Gleason 7 disease at pathological examination (26.2 and 28.2% according to the population defined by Carter et al. and van den Bergh et al., respectively). The significance of the inclusion of Gleason 7 disease in patients selected for AS has recently been addressed. Van den Bergh et al. reported the outcome of 50 patients with biopsy Gleason 7 prostate cancer managed with AS [20]. After a median follow-up of 2.6 years the authors found a 100% cancer-specific survival rate, but the 6-year salvage therapy-free survival was as low as 59%[20]. The authors concluded that AS is an option for patients with Gleason 7 prostate cancer, especially for those with limited life expectancy. However, in patients with long life expectancy the inclusion of Gleason 7 among enter criteria for AS may not represent a safe option.

Several other investigators have addressed the issue of misclassification of patients selected for AS, and several authors proposed to repeat a prostatic biopsy after the initial diagnosis of prostate cancer. Berglund et al.[21] reported the outcome of a repeated biopsy in patients undergoing AS. At the 3-month follow-up biopsy, they reported a 27% rate of upgrading/upstaging. The authors concluded that a repeated biopsy should always be taken to improve the discrimination of who are the best candidates for the AS protocol [21]. In another series of patients managed with AS, Choo et al.[22] reported a 35% Gleason score upgrading at the repeated biopsy, after a median time of 22 months. The present findings corroborate these previous reports for Gleason sum upgrading from Gleason 6 to Gleason 7–10, which was 27.9% and 28.2% when, respectively, the Carter et al. and the van den Bergh et al. criteria were used.

In the Carter et al.[7] series of patients selected for AS, 53 (51%) underwent RP. The authors reported a 20% rate of high-grade or extraprostatic disease, which is also consistent with the present findings. Furthermore, the same group of authors had previously shown a similar rate of incurable (extraprostatic) disease among men in the AS programme undergoing delayed surgical intervention that did not differ from that among men who met the inclusion criteria for the expectant management programme but underwent immediate surgical intervention [12]. Therefore, it is reasonable to conclude that the initial biopsy is unable to correctly assign the Gleason score in about a quarter of patients, as was reported by Chun et al.[23] and Capitanio et al.[24]. Moreover, the evidence of disease progression during the AS programme actually seems to reflect the low ability of prostatic biopsy to correctly stage a patient with prostate cancer. Based on these results, when considering the high rates of treatment-free survival, it is reasonable to speculate that these rates are elevated not only due to the progression of the disease, but instead due to an initial misclassification of the patient characteristics. This finding may have significant implications, as a more accurate staging of the disease should allow higher short-term treatment-free survival rates. As the criteria perform very well in excluding patients with locally advanced or Gleason score ≥8 disease, the major misclassification error applies to patients who may harbour Gleason 7 disease if a RP or a re-biopsy were performed. The identification of these individuals should represent a priority and more accurate models or biomarkers should be developed.

Regarding the number of patients that can be selected for AS, in the present study only 6.9 and 9.7% of patients treated with RP could have been selected for AS according to the criteria proposed by Carter et al. and van den Bergh et al., respectively. These numbers are low, and indicate that few individuals selected for RP harbour prostate cancer that are potentially amenable to AS. In the USA, Barocas et al.[25] analysed the Cancer of the Prostate Strategic Urological Research Endeavor (CaPSURE) database and showed that up to 16% of contemporary American patients with prostate cancer have the clinical characteristics to be enrolled in AS programmes. However, they also showed that only a minority of these patients undergo AS. Carter et al.[7] also reported AS rates as low as 1% among patients with prostate cancer presenting to their institution. These findings underscore the limited applicability and acceptance of AS in patients with clinically localized prostate cancer.

The present study has some limitations. First, our population includes contemporary patients with prostate cancer who elected to undergo RP. It is certain that this surgical cohort is not reflective of all newly diagnosed patients with prostate cancer. Age and comorbidity profiles may be drastically different in the overall cohort of newly diagnosed patients with prostate cancer and a larger proportion of patients may qualify for AS consideration. Indeed, as many as 16% of CaPSURE patients were candidates for AS [25] vs only 6.9–9.7% in the present surgical cohort or 1% at John’s Hopkins University, as reported by Carter et al.[7]. Moreover, the present patients did not undergo a repeat biopsy or even better a staging saturation biopsy. Such measures may substantially reduce the misclassification rate after initial biopsy. Finally, the patient population in the present study originates from a single tertiary care centre. It is possible that Italian patients who are referred to our Institution for RP do not share the same characteristics with other European or American institutions. Therefore, the present findings need to be interpreted considering these limitations. Despite these limitation, the present study is important as it validates the most stringent criteria for AS.

In conclusion, within a surgical cohort of patients with clinically localized prostate cancer, the most stringent criteria for inclusion in AS protocols can identify between 6.9 and 9.7% of patients who fulfil these criteria. Of AS candidates, between 3.3 and 7.1% harbour either ECE, SVI, LNI or Gleason sum 8–10 prostate cancer, which represent clear contraindications to AS. Finally, 27.9–28.2% of AS candidates harbour Gleason 7 prostate cancer, which at most represents a relative contraindication. These findings should be considered in clinical decision-making.

CONFLICT OF INTEREST

None declared.

Ancillary